Rapid Wuchereria bancrofti-Specific Antigen Wb123-Based IgG4Immunoassays as Tools for Surveillance following Mass DrugAdministration Programs on Lymphatic Filariasis

Cathy Steel,a Allison Golden,b Joseph Kubofcik,a Nicole LaRue,b Tala de los Santos,b Gonzalo J. Domingo,b Thomas B. Nutmana

Laboratory of Parasitic Diseases, National Institutes of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland, USAa; Diagnostics Group, PATH,Seattle, Washington, USAb

The Global Programme to Eliminate Lymphatic Filariasis has an urgent need for rapid assays to detect ongoing transmission oflymphatic filariasis (LF) following multiple rounds of mass drug administration (MDA). Current WHO guidelines support usingthe antigen card immunochromatographic test (ICT), which detects active filarial infection but does not detect early exposure toLF. Recent studies found that antibody-based assays better serve this function. In the present study, two tests, a rapid IgG4 en-zyme-linked immunosorbent assay (ELISA) and a lateral-flow strip immunoassay, were developed based on the highly sensitiveand specific Wuchereria bancrofti antigen Wb123. A comparison of W. bancrofti-infected and -uninfected patients (with orwithout other helminth infections) demonstrated that both tests had high sensitivities and specificities (93 and 97% [ELISA] and92 and 96% [strips], respectively). When the W. bancrofti-uninfected group was separated into those with other filarial/hel-minth infections (i.e., onchocerciasis, loiasis, and strongyloidiasis) and those who were parasite uninfected, the specificities ofthe assays varied between 91 and 100%. In addition, the geometric mean response by ELISA of W. bancrofti-infected patientswas significantly higher than the response of those without W. bancrofti infection (P < 0.0001). Furthermore, the Wb123 ELISAand the lateral-flow strips had high positive and negative predictive values, giving valuable information on the size of surveypopulation needed to be reasonably certain whether or not transmission is ongoing. These highly sensitive and specific IgG4tests to the W. bancrofti Wb123 protein give every indication that they will serve as useful tools for post-MDA monitoring.

The Global Programme to Eliminate Lymphatic Filariasis(GPELF), begun in 2000, has led to 952 million treated indi-

viduals, a number that reflects 3.9 billion doses of albendazole pluseither ivermectin or diethylcarbamazine administered in the first11 years of GPELF’s existence (1). Within this time frame, 53 ofthe 73 countries where lymphatic filariasis (LF) is endemic beganimplementation of mass drug administration (MDA), with sev-eral having completed or nearly completed the recommended 5 to6 annual rounds of MDA. This rapid and unprecedented expan-sion of the GPELF has greatly increased the need for surveillancetools to assess interruption of transmission.

The current guidelines established by the WHO for post-MDAsurveillance includes conducting transmission assessment surveys(TAS) that rely on the detection of adult-specific circulating filar-ial antigen (CAg) by use of a lateral-flow immunochromato-graphic test (ICT) (2). While extremely useful in detecting activeWuchereria bancrofti infection, the limitation of the ICT is in itsinability to detect infection prior to the development of adult par-asites, a process that may take up to 18 months following exposureto infective stage larvae. Thus, it has been suggested that a third-stage larva (L3)-specific antibody-based test might be better suitedfor this purpose, given that these antibodies may be detectablesignificantly earlier (months at least) than would the presence ofcirculating Ag, particularly as children 6 to 7 years old are targetedto be the sentinel population for TAS and surveillance (2–4). Tothis end, several serological tools based on antibody to recombi-nant filarial antigens have been tested. These antigens have in-cluded BmR1 and BmSXP (5), WbSXP-1 (6), and Bm33 (7). Per-haps the most widely used antigen in LF surveillance has beenBm14 (8), an antigen that is homologous/identical to BmSXP andWbSXP-1, which has been extensively studied in an enzyme-

linked immunosorbent assay (ELISA) format (4, 8–11). More re-cently, this antigen and several of the others have been comparedusing a multiplex bead assay (3, 12).

Although the sensitivity of the assays based on some of theserecombinant antigens has been excellent in regions of filarial en-demicity in which only Wuchereria and/or Brugia are endemic,there remains the unresolved problem of cross-reactivity in re-gions of co-endemicity for other filarial infections, including Loaloa, Onchocerca volvulus, and Mansonella spp. (e.g., Africa andparts of Central and South America). In a multicenter trial, it wasdemonstrated that both Bm14 and WbSXP were cross-reactive inpatients with either O. volvulus or L. loa infection, and while BmR1was not shown to be cross-reactive, its sensitivity in bancrofti-infected patients is �60% (10). In addition, while antibody toBm33 appears earlier than antibodies to the other recombinantstested, it is a highly immunodominant antigen in the filaria and islikely to be very cross-reactive (12). To address the problem ofspecificity, a recent diagnostic method based on antibodies to theWuchereria bancrofti antigen Wb123 was developed using a lucif-erase immunoprecipitation system (LIPS) (13) that has since beenfurther tested in 2 populations in which W. bancrofti is endemic(12, 14) and has been shown to precede the appearance of anti-

Received 17 April 2013 Returned for modification 20 May 2013Accepted 28 May 2013

Published ahead of print 5 June 2013

Address correspondence to Thomas B. Nutman, tnutman@niaid.nih.gov.

Copyright © 2013, American Society for Microbiology. All Rights Reserved.

doi:10.1128/CVI.00252-13

August 2013 Volume 20 Number 8 Clinical and Vaccine Immunology p. 1155–1161 cvi.asm.org 1155

on February 20, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

genemia. While the LIPS format showed a very high degree ofsensitivity and specificity to Wb123, the expense of new equip-ment needed for this assay (and some of the nonstandard reagentcosts) remains a deterrent to its widespread use.

The purpose of this study was, therefore, to develop rapid,cost-effective formats for the detection of antibodies to the highlysensitive and specific antigen Wb123 for use in the early detectionof filarial transmission following cessation of MDA. First, we de-veloped a rapid immunoassay (ELISA) that can be made availableimmediately to any of the centralized laboratories charged with LFsurveillance. However, since the ELISA still requires the accessi-bility of a laboratory, Wb123 has also been adapted to a lateral-flow strip test that is rapid and can be performed at the point ofcare.

MATERIALS AND METHODSEthics statement. Patient serum samples were collected under severalprotocols approved by the Institutional Review Board of NIAID, with themajority collected under NCT00001230, NCT00342576, or 92-I-0155(inactive). Sample collections as part of larger international field projectswere also approved by the respective governments. Written informedconsent was obtained from all subjects.

Expression and purification of the Wb123-glutathione S-trans-ferase protein. Wb123 was originally identified using Brugia malayi andW. bancrofti L3 expressed sequence tags (ESTs) (13) and has a length of372 amino acids. It most likely belongs to the serpin family of proteins,known to be serine protease inhibitors that are secreted by filaria andhighly immunogenic in humans. Expression and purification of Wb123-glutathione S-transferase (GST) was performed by the Protein ExpressionLaboratory under contract to the National Cancer Institute (FCRF, SAIC,Frederick, MD). Full-length Wb123 was cloned into an expression con-struct suitable for baculovirus expression. This construct, termed GST-tev-Wb123, was a 6.5-kb ampicillin-resistant baculovirus expressionclone, containing a polyhedrin promoter and an N-terminal GST tag. Forexpression and purification, 1 liter of Hi5 cells was infected with baculo-virus containing GST-tev-Wb123 at a multiplicity of infection of 3 andincubated at 21°C for 72 h. Cells were pelleted and then assessed forviability and cell size. The pellet was resuspended in 100 ml of extractionbuffer (20 mM HEPES, 300 mM NaCl, and 2 mM �-mercaptoethanol)supplemented with 1.0 ml of complete protease inhibitor (Sigma-Aldrich,St. Louis, MO) and lysed under high pressure. After microscopic exami-nation of the disrupted cells, 2 U benzonase/ml was added for 20 min onice and the mixture was clarified by centrifugation. The sample was fil-tered and then applied to a 5-ml GSTrap column (Amersham, Pittsburgh,PA). After washing, the protein was eluted from the column over a 20-column volume gradient to 20 mM glutathione in extraction buffer. Theproduct was dialyzed against 1� phosphate-buffered saline (PBS) andanalyzed by SDS-PAGE. For SDS-PAGE, samples were prepared in 1�lithium dodecyl sulfate (LDS) buffer (Bio-Rad, Hercules, CA) with orwithout 25 mM TCEP [tris(2-carboxyethyl)phosphine] reducing agent(Bio-Rad), heated to 95°C, and then loaded onto a 10.5 to 14% precastTris-glycine gel (Bio-Rad). Proteins were visualized with Coomassie bril-liant blue. The final product appeared to be �90% pure and migrated nearits predicted molecular mass of 70.4 kDa (Fig. 1).

Patient samples. Serum samples were chosen from well-defined hel-minth-infected patients and from uninfected patients based on theircountry of origin and/or exposure to Wuchereria bancrofti (Table 1).Themajority of samples were collected as part of studies conducted by theNIAID; samples from Haiti were kindly provided by Patrick Lammie. Foranalysis, a total of 95 patients with W. bancrofti infection were comparedwith subjects without W. bancrofti infection (n � 289 for ELISA analysis;n � 279 for lateral-flow strip analysis), who either had infections withother single filaria (Onchocerca volvulus [n � 116] or Loa loa [n � 40]) orwith soil-transmitted helminths (Strongyloides stercoralis [n � 40]) or

were helminth-uninfected normal subjects from both regions of filarialendemicity (EN) (n � 71 [ELISA] and n � 61 [lateral-flow strips]) andregions of nonendemicity (NEN) (n � 22). Patients with other filarial/helminth infections were diagnosed by skin-snip positivity or by identifi-cation of adult worms in nodules of O. volvulus, positive blood microfi-lariae or eye worm for L. loa, or positive stool examination for S.stercoralis. Parasite-uninfected patients from countries of endemicitywere either from a region free of helminth infection or were excludedbased on negative results for both filarial CAg and by examination forother helminths as discussed above. Samples from patients from regionsof nonendemicity were obtained from normal North American bloodbank donors and other nonexposed individuals.

IgG4 ELISA to Wb123 antigen. To develop a rapid ELISA for use insurveillance post-MDA, conditions were optimized for Wb123-GST-coating concentration, serum dilution, temperature, timing of the incu-bation periods, antibody conjugation and concentration, and develop-ment methodology. The assay was thus finalized as follows. Flat-bottomed96-well plates (Immulon 4; Thermo Scientific, Milford, MA) were coatedwith 10 �g/ml of Wb123-GST in 50 �l/well of PBS overnight at 4°C. Allsubsequent incubations were done at room temperature. Plates werewashed 6 times in PBS-Tween 20 and then blocked for 30 min with 200�l/well of blocking buffer consisting of 1� PBS– 0.05% Tween 20 –5%bovine serum albumin (BSA) (Sigma-Aldrich). Plates were again washedand serum samples were added at 50 �l/well in duplicate at a 1:50 dilutionin a diluent of PBS–1% BSA– 0.05% Tween 20 for 30 min. Added to eachplate was also a positive W. bancrofti serum pool sample (diluted to pro-vide low and high positive controls), a negative serum pool, and ELISAdiluent for plate blanks. Following washing, a 1:5,000 dilution of an alka-line phosphatase-conjugated mouse anti-human IgG4 (clone 6025;Southern Biotech) antibody was added at 50 �l/well in ELISA diluent for30 min. Plates were washed and subsequently developed with 50 �l/well ofalkaline phosphate yellow liquid substrate (pNPP) (Sigma-Aldrich) for 20min. The reaction was stopped with 25 �l/well of 3 N NaOH, and plates

FIG 1 SDS-PAGE of GST-tev-Wb123. Final analysis of the Wb123-GST pro-tein showing migration of the principal protein band (arrow). Samples wereprepared with (�/�) or without (�/�) 25 mM TCEP (reducing agent),heated to 95°C, and loaded onto a 10.5 to 14% precast Tris-glycine gel. Pro-teins were visualized with Coomassie brilliant blue stain. MW, molecular massin kDa.

Steel et al.

1156 cvi.asm.org Clinical and Vaccine Immunology

on February 20, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

were read at 405 nm using a microplate reader. It should be noted that thisassay also works well using the horseradish peroxidase (HRP)-conjugatedversion of the antibody and an HRP developer such as ABTS [2,2=-azino-bis(3-ethylbenzthiazoline sulfonic acid)]; however, all the results in thisstudy are based on the alkaline phosphatase assay.

To simplify the process further and shorten the assay time for investi-gators doing surveillance of large populations, Immulon 4 plates werecoated and blocked as above followed by a final wash. Plates were tappedto remove excess liquid and then allowed to air dry for several hours atroom temperature. When dry, plates were sealed with an absorbent silicagel pack in a sealer pouch (Sorbent Systems, Los Angeles, CA) and storedat 4°C. Sealed plates have thus far been tested after storage for 6 weeks andproved comparable to plates freshly coated.

Comparison of Wb123 ELISA with other assays. To compare the newWb123-GST IgG4 ELISA results with those of antibody assays used inprevious studies, either using Wb123 or whole filarial antigen, correla-tions were made for a subset of patients between the ELISA results fromthis study with the results from 3 other platforms as described below.

Luciferase immunoprecipitation system. The LIPS assay for detec-tion of both IgG and IgG4 antibodies to Wb123-GST has been describedin detail previously (13). Very briefly, sera were diluted 1:10 and added to

a Ruc-Wb123 enzyme reporter. Following a short incubation of 5 min, asuspension of either protein A/G beads (for IgG analysis) or anti-IgG4beads was added. After a second 5-min incubation, plates were washedand processed on a Berthold LB 960 Centro microplate luminometer us-ing a coelenterazine substrate mix (Promega, Madison, WI); data are ex-pressed in luminometer units.

Suspension array technology for IgG4. The Luminex suspension ar-ray technology assay has been described in detail elsewhere (15). For thisanalysis, 160 �g of Wb123-GST was coupled to 16 � 106 activated Sero-MAP beads (Luminex Corporation, Austin, TX). Coupled beads wereadded to a 1:1,000 dilution of patient sera at a concentration of 5,000beads/well in a prewetted filter plate (EMD Millipore, Billerica, MA).Following a 2-h incubation on a plate shaker in the dark, plates werewashed through a vacuum manifold. Biotinylated mouse anti-humanIgG4 (clone 6025; Hybridoma Reagent Laboratory, Baltimore, MD) wasthen added at a 1:1,000 dilution for 1 h with shaking (in the dark) afterwhich streptavidin-conjugated phycoerythrin (PE) (Jackson Immuno-Research, West Grove, PA) was added at 2 �g/ml for an additional 30 min.Plates were washed and samples were read using the Bio-Plex System(Luminex); data are expressed as mean fluorescence intensities (MFI).

ELISA for IgG4 to adult filarial antigen. IgG4 antibody to a salineextract of Brugia malayi adult antigen (BmA) was measured as previouslydescribed (16).

Wb123 lateral-flow strip immunoassay. Lateral-flow strip materialswere assembled on Lohmann diagnostic strip backing (300 by 75 mm) ina card format, using 25-mm-wide nitrocellulose membrane, 22-mm-wideconjugate pad material, and 32-mm-wide absorbent wick material (Fig. 2)and cut into 4-mm-wide strips. A BioDot XYZ reagent dispenser was usedto apply Wb123 antigen and control antibody (goat anti-mouse; JacksonImmunoResearch) stripes onto the nitrocellulose membrane and to sprayanti-human IgG4 antibody (Hybridoma Reagent Laboratory) conjugatedto gold colloid onto the conjugate pad material. For detection of theWb123 antibody, 2.5 �l of patient serum or plasma was added to thecenter of the test strip approximately 4 mm from the conjugate pad (Fig.2). The conjugate end of the strip was subsequently placed in a well of aflat-bottomed 96-well plate with 100 �l of lateral-flow test running buffer.Strips were read at 20 min, removed from the well, and then read againwhen dry.

Statistical analyses. For calculation of the IgG4 antibody response toWb123-GST by ELISA, optical densities (ODs) of duplicate patient sam-ples were averaged and then divided by the OD of the corresponding plateblank. Determination of the positive threshold to Wb123 was accom-plished using receiver operator characteristic (ROC) curves (GraphPadPrism version 6.0) to compare the ratios of W. bancrofti-infected patientsto all patients not infected with W. bancrofti, including both helminth-infected and parasite-uninfected patients from regions of W. bancroftiendemicity and nonendemicity. To detect differences between the ratiosof patients with and those without W. bancrofti infection, the Mann-Whitney test was used. Correlations between the Wb123 IgG4 ELISA andother immunological assays were accomplished using the Spearman ranktest (GraphPad Prism). A sensitivity and specificity calculator was utilizedto determine positive and negative predictive values for both the ELISAand the lateral-flow strips.

RESULTSELISA results. The ratios of the ODs of patient samples to theplate blanks were calculated for both patients with W. bancroftiinfection and for each group of patients without W. bancrofti in-fection (Fig. 3). The ratios obtained using serum samples frompatients with W. bancrofti infection were significantly higher thanthose obtained using serum samples from the entire group of in-dividuals without W. bancrofti infection (P � 0.0001), with strik-ing differences in the geometric means (GM) between the individ-ual groups compared with W. bancrofti-infected patients (GM ofthose with W. bancrofti infection, 16.44; with O. volvulus, 1.48;

TABLE 1 Patient populations

Patient infection status and country of originNo.(n � 384)

Wuchereria bancrofti positive 95Cook Islands 59India 18Haiti 16Guyana 2

Onchocerca volvulus positive 116Ecuador 46Guatemala 41Ghana 3Cameroon 3Nigeria 1Expatriates 22

Loa loa positive 40Cameroon 1Expatriates (Cameroon, CAR, DRC, Benin, Gabon) 39

Strongyloides stercoralis positive 40Jamaica 1Central America 12West Africa 1Southeast Asia 18Middle East 1Expatriates (Southeast Asia, Central and South

America)7

Non-helminth infected from country where LF isendemic

71a

Ecuador (Quito) 20Guatemala 17Haiti (nonexposed) 16Mali (Bamako) 18

Noninfected from country of W. bancroftinonendemicity

North America 22a n � 384 represents the number of patients tested by ELISA. For strip testing, n � 374(Haiti nonexposed, n � 6 rather than 16).

Rapid Wb123-Based Immunoassays for LF Surveillance

August 2013 Volume 20 Number 8 cvi.asm.org 1157

on February 20, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

with L. loa, 1.08; with S. stercoralis, 1.14; in EN, 0.99; in NEN,0.87).

To determine the positive threshold response to Wb123, theratios of W. bancrofti-infected patients were compared to allnon-W. bancrofti-infected patients in a ROC analysis (Fig. 3 inset).By defining the optimal balance between sensitivity and specific-ity, we determined the positive threshold cutoff to be a sample/plate blank ratio of 2.5, giving a sensitivity of 92 to 93% and aspecificity of 96 to 97%. Using this threshold, 7 patients had false-negative results (GM response, 1.24) and 9 had false-positive re-sults (GM, 9.48) (Fig. 3). Interestingly, all of the false positiveswere from O. volvulus-infected patients from Guatemala (n � 6)or Ecuador (n � 3), regions where bancroftian filariasis is not

endemic. The 7 patients with false-negative results (all of whomwere microfilaria positive) were distributed among India (n � 2),the Cook Islands (n � 1), Guyana (n � 1), and Haiti (n � 3).

To further test the performance of the Wb123 IgG4 ELISA, thepositive and negative predictive values for the assay were deter-mined for each clinical group and for all W. bancrofti-negativepatients (including those with other non-W. bancrofti helminthinfections as well as parasite-uninfected individuals) compared toW. bancrofti-positive patients (Table 2). The sensitivity of the as-say as established by these calculations was 92.6%, with specificityranging from 92.2% (for O. volvulus patients) to 100% for “allparasite uninfected” and for patients with either loiasis or strongy-loidiasis. Overall, based on calculations using the W. bancrofti-negative group, the specificity of the test was determined to benearly 97%. The ELISA also gave high positive (ranging from 90.7to 100%) and negative (ranging from 85.1 to 97.6%) predictivevalues. The lower negative predictive values for L. loa and S. ster-coralis patients were most likely due to the smaller numbers ofsamples in these patient groups (n � 40 for each).

To determine how the Wb123 IgG4 rapid ELISA compareswith platforms used in other studies (12–14, 17), subsets of sam-ples previously tested by these methods in this laboratory werecompared with the current ELISA results. The primary compari-son was to the recently described Wb123-GST LIPS assay (13).Approximately 35% (134/384) of the patients used for the ELISA-based test had previously been tested by the IgG-based LIPS assay.Despite the differences in the isotypes measured between the LIPSand the ELISA, the tests were strongly correlated (r � 0.747, P �0.0001) (Fig. 4). In addition, a much smaller subset (n � 16) wascompared to the LIPS IgG4 assay, also showing a strong positiverelationship (n � 0.772, P � 0.0008).

Prior to the use of recombinant proteins, an IgG4-based ELISAto the adult filarial antigen BmA was frequently used for diagnosis(16). Interestingly, while Wb123 is primarily expressed in L3s, theIgG4 antibody levels to Wb123 were positively correlated (r �0.369, P � 0.0040) to the levels of BmA-specific IgG4 for thesubset of patients tested (n � 59), though understandably the

FIG 3 Determination of the appropriate cutoffs for the Wb123 ELISA-basedimmunoassay. The ODs of the duplicate patient samples were averaged andthen divided by the corresponding plate blanks. Each dot represents an indi-vidual patient sample from patients with W. bancrofti (shown in red), O. vol-vulus (in blue), and other infections (in black). Shown also are parasite-unin-fected controls from regions of filaria endemicity and nonendemicity (also inblack). The geometric means (GM) for each group are illustrated by the indi-vidual horizontal lines and the normal range is indicated by the gray box. Theinset shows the ROC curve for specificity/sensitivity based on the sample/plateblank ratios of 95 W. bancrofti-infected patients compared to 289 non-W.bancrofti-infected patients.

FIG 2 Representation of Wb123 lateral-flow strips. (A) Diagrammatic illustration of the lateral-flow strips showing the placement of the antibody conjugatesand the appearance of the detection lines before (top) and following (bottom) the addition of patient sera. (B) Negative (top) and positive (bottom) results fromrepresentative patient serum samples.

Steel et al.

1158 cvi.asm.org Clinical and Vaccine Immunology

on February 20, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

relationship was not as strong as for the LIPS assay to Wb123. IgG4antibody levels to Wb123 were not, however, correlated to theantibody levels to a Brugia microfilarial antigen preparation (datanot shown). Finally, because a multiplex bead analysis had re-cently been used to compare several different recombinant filarialproteins (12), IgG4 responses in a small group of 15 patients werecompared using both the Wb123 ELISA and Wb123 Luminexplatforms. As shown for the LIPS assay, the results of the ELISAand Luminex formats were strongly correlated (r � 0.833, P �0.0003) (Fig. 4).

Lateral-flow strip immunoassay. Lateral-flow strips were readboth at 20 min and after drying. Importantly, strips displayedgreat stability between the initial reading and that done subse-quently, with only 4/374 (1.1%) changing from negative to posi-tive status following drying. The sensitivity and specificity calcu-lations were comparable to those derived from the ELISA (Table2). Strips read at 20 min showed a sensitivity of 91.6% with spec-ificity ranging from 92.2% (when patients with onchocerciasiswere used as the comparator) to 100% (for Loa and Strongyloidespatients) with an overall test specificity of 96.4%. The dry strips

showed slightly greater sensitivity (92.6%) but also marginallylower specificity (95.3% overall, with a range for individual pa-tient groups of 90.5 to 100%). The positive and negative predictivevalues were also similar to those obtained for the ELISA, withoverall values of 89.7% and 87.1% (positive predictive values) and97.1% and 97.4% (negative predictive values) for the 20-min anddry reads, respectively. Overall, there was quite good consistencybetween the ELISA results and those of the lateral-flow strips forpatient status. In an examination of just the results that differed inoutcome between the 20 min strips and the ELISA, 14/374 (3.7%)of sera tested gave a discordant result (O. volvulus, strips, 5 falsepositives [FP] and ELISA, 5 FP; W. bancrofti, strips, 2 false nega-tives [FN] and ELISA, 1 FN; EN, strips, 1 FP). Comparison be-tween the ELISA and the dry strips showed a 4% (15/374) discor-dance (O. volvulus, strips, 6 FP, and ELISA, 4 FP; W. bancrofti,strips, 2 FN, and ELISA, 2 FN; EN, strips, 1 FP).

DISCUSSION

One of the immediate needs in the post-MDA world is the re-quirement for a highly sensitive and specific test for the assessment

TABLE 2 Sensitivities and specificities of Wb123 IgG4 ELISA and lateral-flow strips

W. bancrofti infectedvs indicatedinfection

IgG4 strip results (%)

IgG4 ELISA results (%) (20-min read) Dry

Sensitivity Specificity

Positivepredictivevalue

Negativepredictivevalue Sensitivity Specificity

Positivepredictivevalue

Negativepredictivevalue Sensitivity Specificity

Positivepredictivevalue

Negativepredictivevalue

Parasite uninfected 92.6 100 100 93.0 91.6 98.8 98.9 91.1 92.6 97.6 97.8 92.1O. volvulus 92.6 92.2 90.7 93.9 91.6 92.2 90.6 93.0 92.6 90.5 88.9 93.8L. loa 92.6 100 100 85.1 91.6 100 100 83.3 92.6 100 100 85.1S. stercoralis 92.6 100 100 85.1 91.6 100 100 83.3 92.6 100 100 85.1All non-W. bancrofti

infected92.6 96.9 90.7 97.6 91.6 96.4 89.7 97.1 92.6 95.3 87.1 97.4

FIG 4 Correlations between the Wb123 ELISA and other immunoassay formats. In a subset of patients, the ratios of patient samples to the corresponding plateblanks in the Wb123 IgG4 ELISA were compared to the Wb123 assays by LIPS (IgG [upper left], IgG4 [upper right]), the IgG4 Luminex assay (lower right), andthe IgG4 ELISA antibody response to adult filarial antigen (BmA; lower left). The Spearman rank correlation analysis was used to generate r and P values.

Rapid Wb123-Based Immunoassays for LF Surveillance

August 2013 Volume 20 Number 8 cvi.asm.org 1159

on February 20, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

of potential ongoing transmission of the parasites that cause LF. Inaddition, such a test should be fairly rapid and relatively inexpen-sive to perform on large numbers of samples in centralized labo-ratories or ideally at the point of care. Based on WHO guidelines,the current tool for assessment of Wuchereria bancrofti transmis-sion is the ICT circulating antigen card test (2). However, thefocus on surveillance tools has shifted toward the use of antibody-based assays that can detect exposure to L3s, allowing for earlywarning signals of potential W. bancrofti transmission. Many suchantibody tests have already been developed (5–8) but the require-ment for specificity in regions of coendemicity for other filarialinfections (e.g., Africa, parts of the Americas) has not been met.The purpose of this study was to adapt a recombinant W. bancroftiantigen, Wb123, already proven to be sensitive and specific inLIPS assays (13), to more standard immunoassay formatting, in-cluding those that could be used at the point of care. As found inthe previous LIPS assay based on Wb123 (13), both the IgG4ELISA and the lateral-flow strips demonstrated striking specificityand sensitivity when comparing patients infected with W. ban-crofti to those infected with O. volvulus, L. loa, or S. stercoralis,parasites with highly homologous genomes and with potentialantigenic cross-reactivity. As expected, individuals without hel-minth infection living in regions of either filarial endemicity ornonendemicity showed no IgG4 reactivity to Wb123 for the ELISAand only very minimal cross-reactivity by strip analysis.

The use of the Wb123 LIPS test for surveillance, particularly inchildren, was recently demonstrated in a South Pacific populationthat had undergone a single island-wide treatment with diethyl-carbamazine (14). However, this study was based on a populationwhere LF transmission had decreased but had not been elimi-nated. Nevertheless, when the current ELISA platform was usedwith dried blood spots, no reactivity was seen in individuals (n �292, age � 7 years; n � 45, age 7 years) from 6 Malian villagesthat had undergone 6 rounds of MDA (data not shown). In addi-tion to post-MDA surveillance, these rapid diagnostic assays willbe highly useful in detecting new infections following an influx ofimmigrants from a region of filarial endemicity to either a regionof endemicity or to one cleared of infection following MDA.

That the Wb123 ELISA and the lateral-flow strips have highpositive and negative predictive values will be beneficial in theanalysis of large posttreatment populations, giving needed infor-mation on the necessary sampling size to adequately survey a pop-ulation. Moreover, these tests should be of benefit in mappingareas in which the W. bancrofti status is unknown (e.g., parts ofCameroon), given their relatively high throughput, reduced cost,and rapid turnaround time.

Additionally, for the ELISA, the results correlated well withother assay formats (Fig. 4), suggesting the consistency of thisplatform for immediate expanded use. Furthermore, we have alsotested the use of Wb123 precoated and blocked plates and foundthem quite comparable to freshly coated plates (data not shown),further simplifying the transfer of this assay to regions of the worldwhere needed. Since most regional and even district laboratorieshave the ability to perform ELISAs, we developed a methodologythat requires no other equipment (such as incubators or auto-mated washers) and works well at ambient temperatures.

The ultimate goal of any diagnostic is to be as rapid and accu-rate as possible. The highly sensitive and specific IgG4-basedELISA is one step in that direction. Indeed, the ongoing develop-ment and finalization of Wb123 in a lateral-strip format are the

next logical steps in making a field-accessible, point-of-care test.Thus far, these strips have shown remarkable consistency with theELISA and comparable sensitivity and specificity in positive andnegative antibody determination. In addition, a similar prototypeof this lateral-flow format has already been successfully tested withwhole blood rather than sera for detection of antibody to the on-chocerciasis antigen Ov16 (A. Golden, submitted for publication).For these strips (n � 99), there were only 3 results that were dis-cordant between the strips assayed against sera or plasma (1 falsepositive) and those with whole blood (1 false negative and 1 falsepositive). However, until this rapid point-of-care test is fully de-veloped and commercialized for Wb123, the IgG4 ELISA pre-sented in this study is both ready to use and inexpensive (on a persample basis) and should serve as a highly valuable tool for pop-ulations undergoing postcontrol surveillance currently and in thevery near future.

ACKNOWLEDGMENTS

We thank Dominic Esposito, William Gillette, and other members of theSAIC Protein Expression Laboratory for the expression and purificationof Wb123-GST.

This research was supported by the Intramural Research Program ofthe NIH, NIAID.

REFERENCES1. World Health Organization. 2012. Global programme to eliminate lym-

phatic filariasis: progress report, 2011. Weekly epidemiological record, no.37, 87, 346 –356. World Health Organization, Geneva, Switzerland http://www.who.int/wer/2012/wer8737.pdf.

2. World Health Organization. 2011. Global programme to eliminate lym-phatic filariasis: monitoring and epidemiological assessment of mass drugadministration. World Health Organization, Geneva, Switzerland.

3. Moss DM, Priest JW, Boyd A, Weinkopff T, Kucerova Z, Beach MJ,Lammie PJ. 2011. Multiplex bead assay for serum samples from childrenin Haiti enrolled in a drug study for the treatment of lymphatic filariasis.Am. J. Trop. Med. Hyg. 85:229 –237.

4. Ramzy RM, El Setouhy M, Helmy H, Ahmed ES, Abd Elaziz KM, FaridHA, Shannon WD, Weil GJ. 2006. Effect of yearly mass drug adminis-tration with diethylcarbamazine and albendazole on bancroftian filariasisin Egypt: a comprehensive assessment. Lancet 367:992–999.

5. Abdul Rahman R, Hwen-Yee C, Noordin R. 2007. Pan LF-ELISA usingBmR1 and BmSXP recombinant antigens for detection of lymphatic filar-iasis. Filaria J. 6:10. doi:10.1186/1475-2883-6-10.

6. Pandiaraja P, Arunkumar C, Hoti SL, Rao DN, Kaliraj P. 2010. Evalu-ation of synthetic peptides of WbSXP-1 for the diagnosis of human lym-phatic filariasis. Diagn. Microbiol. Infect. Dis. 68:410 – 415.

7. Krushna NS, Shiny C, Dharanya S, Sindhu A, Aishwarya S, Narayanan RB.2009. Immunolocalization and serum antibody responses to Brugia malayipepsin inhibitor homolog (Bm-33). Microbiol. Immunol. 53:173–183.

8. Weil GJ, Curtis KC, Fischer PU, Won KY, Lammie PJ, Joseph H,Melrose WD, Brattig NW. 2010. A multicenter evaluation of a newantibody test kit for lymphatic filariasis employing recombinant Brugiamalayi antigen Bm-14. Acta Trop. 120:S19 –S22.

9. Helmy H, Weil GJ, Ellethy AS, Ahmed ES, Setouhy ME, Ramzy RM.2006. Bancroftian filariasis: effect of repeated treatment with diethylcar-bamazine and albendazole on microfilaraemia, antigenaemia and antifi-larial antibodies. Trans. R. Soc. Trop. Med. Hyg. 100:656 – 662.

10. Lammie PJ, Weil G, Noordin R, Kaliraj P, Steel C, Goodman D,Lakshmikanthan VB, Ottesen E. 2004. Recombinant antigen-based an-tibody assays for the diagnosis and surveillance of lymphatic filariasis—amulticenter trial. Filaria J. 3:9. doi:10.1186/1475-2883-3-9.

11. Weil GJ, Kastens W, Susapu M, Laney SJ, Williams SA, King CL,Kazura JW, Bockarie MJ. 2008. The impact of repeated rounds of massdrug administration with diethylcarbamazine plus albendazole on ban-croftian filariasis in Papua New Guinea. PLoS Negl. Trop. Dis. 2:e344.doi:10.1371/journal.pntd.0000344.

12. Hamlin KL, Moss DM, Priest JW, Roberts J, Kubofcik J, Gass K, StreitTG, Nutman TB, Eberhard ML, Lammie PJ. 2012. Longitudinal moni-

Steel et al.

1160 cvi.asm.org Clinical and Vaccine Immunology

on February 20, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from

toring of the development of antifilarial antibodies and acquisition ofWuchereria bancrofti in a highly endemic area of Haiti. PLoS Negl. Trop.Dis. 6:e1941. doi:10.1371/journal.pntd.0001941.

13. Kubofcik J, Fink DL, Nutman TB. 2012. Identification of Wb123 as anearly and specific marker of Wuchereria bancrofti infection. PLoS Negl.Trop. Dis. 6:e1930. doi:10.1371/journal.pntd.0001930.

14. Steel C, Kubofcik J, Ottesen EA, Nutman TB. 2012. Antibody to thefilarial antigen Wb123 reflects reduced transmission and decreased expo-sure in children born following single mass drug administration (MDA).PLoS Negl. Trop. Dis. 6:e1940. doi:10.1371/journal.pntd.0001940.

15. Fouda GG, Leke RF, Long C, Druilhe P, Zhou A, Taylor DW, JohnsonAH. 2006. Multiplex assay for simultaneous measurement of antibodies tomultiple Plasmodium falciparum antigens. Clin. Vaccine Immunol. 13:1307–1313.

16. Lal RB, Ottesen EA. 1988. Enhanced diagnostic specificity in humanfilariasis by IgG4 antibody assessment. J. Infect. Dis. 158:1034 –1037.

17. Steel C, Ottesen EA, Weller PF, Nutman TB. 2001. Worm burden andhost responsiveness in Wuchereria bancrofti infection: use of antigen de-tection to refine linear assessments from the South Pacific. Am. J. Trop.Med. Hyg. 65:498 –503.

Rapid Wb123-Based Immunoassays for LF Surveillance

August 2013 Volume 20 Number 8 cvi.asm.org 1161

on February 20, 2020 by guest

http://cvi.asm.org/

Dow

nloaded from